A Minor Multifault Cooperative Diagnosis Method for Lithium-Ion Power Batteries Based on Multi-Information Fusion

Fault diagnosis is a key technology to ensure the safe operation of power batteries. However, batteries minor faults are characterized by diverse triggers and coupled features, which are easily covered by noise. Early diagnosis of minor faults is extremely challenging. The traditional threshold method is difficult to determine the optimal threshold, which cannot diagnose minor faults. Therefore, a minor multifault cooperative diagnosis framework based on multi-information fusion is proposed. On the one hand, the variance generalized S-transform is presented to detect minor faults. Further, internal short circuits, external short circuits, and poor contact faults are distinguished based on the voltage pseudo-variance and temperature rise. On the other hand, the variable scale distance integral distance and voltage spectrum integration method are invented to diagnose self-discharge faults and sensor faults, respectively. The equivalent internal resistance approach is used to detect inconsistent faults. The experimental results demonstrate that the presented approach achieves a 98.5% fault detection rate and a 97.7% detection accuracy rate. The detection results of different types of batteries faults and simultaneous faults of diverse cells prove the robustness of the proposed technology.